Deep Snow: Synthesizing Remote Sensing Imagery with Generative Adversarial Nets
This work addresses the challenge of synthesizing remote sensing data for applications like environmental monitoring, but it is incremental as it builds on existing GAN methods.
The paper tackles the problem of generating realistic pervasive changes in remote sensing imagery using generative adversarial networks (GANs) in an unpaired training setting, achieving results where generated images appear perceptually similar to real ones but with identifiable artifacts.
In this work we demonstrate that generative adversarial networks (GANs) can be used to generate realistic pervasive changes in remote sensing imagery, even in an unpaired training setting. We investigate some transformation quality metrics based on deep embedding of the generated and real images which enable visualization and understanding of the training dynamics of the GAN, and may provide a useful measure in terms of quantifying how distinguishable the generated images are from real images. We also identify some artifacts introduced by the GAN in the generated images, which are likely to contribute to the differences seen between the real and generated samples in the deep embedding feature space even in cases where the real and generated samples appear perceptually similar.